** Note: ** A model should be trained before using the app for seizure detection
- Launch Conda Shell Prompt, navigate to seizy_ml directory and activate the virtual environment.
cd ./seizy_ml
conda activate seizyml
- Set path for data processing.
python cli.py setpath 'path'
- This is the parent path where the directory ('data_dir') with h5 data resides configuration settings.
- All subsequent folders and model predictions will reside here.
- Run file check.
python cli.py filecheck
⚠️ This step checks that the h5 files have the correct dimensions. For help on how to convert files to h5 have a look at the h5_conversion script.
- Preprocess data.
- This is the step where the h5 data files will be filtered and large outliers will be removed.
python cli.py preprocess
- Generate model prections.
python cli.py predict
- Here selected features will be extracted and model predictions will be generated using the selected model from training.
- Verify seizures and adjust seizure boundaries.
- This will launch a prompt to allow for file selection for verification.
- After the file selection, a GUI will be launched for seizure verfication and seizure boundary adjustment.
python cli.py verify
- Get seizure properties. -This step will generate a csv file with seizure properties for each h5 file.
python cli.py extractproperties